Question 8

Unclassified

Which technique is most appropriate to handle a categorical feature with thousands of unique values (high cardinality)?